Objective: To investigate, in patients with severe septic shock, the adrenocortical function assessed by daily plasma cortisol determinations during the first 72 h and by the short synthetic ACTH stimulation test performed within 24 h of the onset of shock. Design: Prospective clinical investigation. Setting: Medical intensive care unit in a university teaching hospital. Patients: 40 consecutive patients with documented septic shock requiring at least hemodynamic resuscitation and respiratory support. Interventions: There were no interventions. Measurements and results: Basal cortisol concentrations were increased with a mean value of 36.8 mug/dl (range 7.9 - 113). Of the overall cortisol determinations 92% were above 15 mug/dl. No statistically significant differences in basal cortisol concentrations were found when survival, type of infection, and positive blood cultures were considered. Patients with hepatic disease had significantly higher cortisol (50.1(+/- 6.2) mug/dl versus 35.9(+/- 3.3) mug/dl, p = 0.035) levels compared to other patients. No correlations were found between basal plasma cortisol concentrations and factors such as SAPS, OSF, hemodynamic measurements, duration of shock, and amount of vasopressor and/or inotropic agents. Cortisol concentrations and significant but weak correlation with ACTH levels in survivors (r = 0.4; p = 0.03; n = 28) but not in non-survivors (r = 0.03; p = 0.85; n = 52). Cortisol levels in non-survivors increased significantly from enrollment time to the 72nd hour of the survey (day 1: 38.9(+/-3.8)mug/dl versus day 3: 66.7(+/-17.1) mug/dl; p = 0.046) and were significantly higher than those recorded in survivors. Responses to the short ACTH stimulation test were not significantly different between survivors and non-survivors. According to the different criteria used to interpret the response to the ACTH stimulation test, incidence of adrenocortical insufficiency was highly variable ranging from 6.25 - 75% in patients with septic shock. Only one patient had absolute adrenocortical insufficiency (basal cortisol level below 10 mug/dl; response to the ACTH stimulation test below 18 mug/dl). Conclusion: Our data suggest that in a selected population of patients with severe septic shock single plasma cortisol determination has no predictive value. The short ACTH stimulation test performed within the first 24 h of onset shock can neither predict outcome nor estimate impairment in adrenocortical function in patients with high basal cortisol level. Adrenal insufficiency is rare in septic shock and should be suspected when cortisol level is below 15 mug/dl and then confirmed by a peak cortisol level lower than 18 mug/dl during the short ACTH stimulation test.
